Review TRENDS in Plant Science Vol.10 No.11 November 2005

Evolution of diversity: the Cape flora

H. Peter Linder

Institute for Systematic Botany, University of Zurich, Zollikerstrasse 107, Zurich CH-8008, Switzerland

Although the environmental correlates of plant species the species are endemic to this region [9]. This high level richness have long received attention, research into the of endemism, the presence of distinctive Cape floral clades genesis of this diversity is in its infancy. The recent [12] and a large variation in the richness of the clades, development of molecular dating methods and the makes this a suitable flora to use as a model system for an inference of robust phylogenetic hypotheses have investigation into the genesis of plant diversity. made it possible to investigate this problem. I use the Cape flora as a model to show that much modern Assembly of the flora diversity could be the result of recruiting diverse For many years, the Cape flora was thought to be lineages over the entire Cenozoic into this flora, composed of three components. The ‘Antarctic’ or Gond- followed by in situ diversification (thus increasing wanan component was postulated to be a relict of the species richness) of at least some of these lineages Cretaceous Gondwanan flora [13]; possible elements are over a long period in an environmentally heterogeneous the gymnosperms Podocarpus and Widdringtonia, as well area. as Proteaceae and Restionaceae [9]. The African com- ponent, which makes up the bulk of the Cape flora, was shared with tropical Africa, leading to the interpretation Cape flora of the Cape flora as a specialized African flora [14]. The The global distribution of plant diversity and endemism is Eurasian component might have migrated relatively uneven, with concentrations of species in the Andes, the recently southwards along the African mountains from Atlantic coastal forests of Brazil, the southern tip of Eurasia [15]. This vicariance view (which seeks to Africa, the East Indies and the eastern Himalayas [1]. understand biogeography as the result of fragmentation Species richness has been correlated, among other things, of older distribution ranges [16]) sees the Cape flora as with palaeoclimatic stability [2] and the combination of composed predominantly of fragments of two once more energy levels and water availability [3], which are all widespread floras: a Cretaceous Gondwanan flora and a variously correlated with the latitudinal gradient [4]. Tertiary African flora. However, only with the development of robust phyloge- However, phylogenetic analyses (Box 1) in the past netic and molecular dating methods in the past five years decade have revealed a much more complex picture. There has it become possible to address the evolutionary genesis are several post-Gondwanan intercontinental relation- of this diversity. Species diversity is, in many floras, the ships: for example, Prionium with its sister Thurnia in result of a few species-rich lineages, whereas most other South America [17], and Pelargonium with the Peristera lineages include only a few species [5]. The central section in Australia [18]. A recent, detailed inventory of question is when, and why, the evolution of these the affinities of the Cape flora showed that it contains species-flocks started. lineages with closest relatives on all continents but that Understanding the genesis of plant diversity is the most common relationships lie with the Australian important for long-term conservation planning: only by flora [19]. The earliest recruitment of an angiosperm knowing the evolutionary responses of plants to past lineage into the Cape flora has been dated to the environmental changes can the consequences of future Cretaceous. Since then there has been a regular incorpor- changes be predicted. Conservation planning should ation of ever more lineages into this flora [19], with incorporate models of the past diversification of the local migration into the flora not blocked by the opening of the floras [6–8]. southern oceans. Oceans and geographical distance limit The flora of the Cape Floristic Region, which comprises migration in plants (as is evident from the problems with 2 w9000 species in an area of 90 000 km , has a species invasive plants transported by humans across diverse richness comparable to neotropical floras [9,10] and is barriers) but do not present absolute barriers. substantially richer than the other Mediterranean-type Similar shifts in interpretation have occurred for other ecosystems (e.g. Western Australia, Chile, California and floras. In the past, the flora of New Zealand [16] and the Mediterranean basin) [11]. The Cape flora can readily Australia [20] were interpreted as fragments of the be defined by an unusual combination of families: the eastern Gondwanan flora, which included South America, Proteaceae, Restionaceae, Ericaceae and Aizoaceae are Antarctica, New Zealand and Australia. The flora of South among the most diverse and are also the ecologically America was seen as a mixture of the southern Gondwa- dominant families (Figure 1). Furthermore, 67% of nan flora and a northern tropical flora [21], itself a

Corresponding author: Linder, H.P. ([email protected]). fragment of the South American–African–Indian (West Available online 6 October 2005 Gondwanan) tropical flora [22]. However, more recent www.sciencedirect.com 1360-1385/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.tplants.2005.09.006 Review TRENDS in Plant Science Vol.10 No.11 November 2005 537

Figure 1. Selected species of typical Cape floral families: (a) Conophytum speciosus (Aizoaceae), (b) Phylica plumigera (Rhamnaceae), (c) Erica glauca (Ericaceae), (d) Tritoniopsis triticea (Iridaceae), (e) Mimetes arborea (Proteaceae), (f) Restio strobilifer (Restionaceae), (g) Heliophila sp. (Brassicaceae). research on the flora of New Zealand suggests that much more dynamic biogeographical worldview that does not of the flora was assembled by long-distance dispersal [23] assign whole floras to categories but rather seeks to and that most plant lineages arrived in New Zealand understand the evolutionary history of each of its during the Neogene, mostly from Australia. A surprising constituent clades. number of lineages in the Australian flora have been recruited from across the oceans [24]. Similarly, the South Age of the modern species flocks American flora had received immigrants throughout the Half of the species richness of the Cape flora can be traced Cenozoic from Africa, Laurasia and Australia by long- to 30 radiations [12]. This is similar to the Hawaiian flora, distance dispersal and stepping-stone dispersal; these where half of the richness is derived from 20 radiations immigrants make up more than 10% of the lowland [29], and the Australian flora, where the 10 largest tropical rainforest flora [25]. families comprise 50% of the plant species of Australia The Cape flora is thus not unusual in its ability to [30]. With the advent of molecular dating methods it has recruit lineages from all continents. Consequently, it can become possible to estimate the start of each of these be confusing to refer to it as a Gondwanan or an African radiations. The first published study, which was on the flora. However, individual clades can still be associated Rhamnaceae genus Phylica, indicated a Late Miocene with an area where the clade initially or primarily initiation date [31]. However, the investigation of diversified. Proteaceae or Restionaceae could still be additional Cape lineages returned mostly older dates: referred to as ‘Gondwanan’ without implying that the Pelargonium [32] and Indigofera [33] date from the middle flora in which they are found is Gondwanan. So we can Miocene; Irideae, Ixioideae and Nivenioideae (Iridaceae) also recognize Cape floral clades (e.g. the African from Oligocene or early Miocene [34]; and the African Restionaceae and the tribe Proteae); these are indeed Restionaceae radiation was probably initiated in the widespread on many tropical African mountains [26–28] Oligocene [35]. At the other extreme, the analysis of and also in Australia [18] and New Zealand. This reflects a Heliophila suggested an initiation of the radiation in www.sciencedirect.com 538 Review TRENDS in Plant Science Vol.10 No.11 November 2005

the Pliocene [36]. This implies that some lineages in the Box 1. Inferring evolutionary history without fossils Cape flora have been accumulating species over the past Phylogenetic hypotheses specify the genealogical relationships 35 million years (My), whereas others started this process among species. Two recent developments have added much rigour much more recently. to the construction of phylogenetic hypotheses. The development of cladistic analytical methods provided a logically rigorous method for The radiation of Australian Restionaceae started extracting a phylogenetic signal from comparative data and the substantially earlier (in the Eocene) than that of the development of the polymerase chain reaction and automatic African Restionaceae [37]. Eocene to early Oligocene dates sequencers has provided access to the almost unlimited data locked were also obtained for the initiation of the radiations of the up in DNA molecules [62]. Australian Eucalyptus, Casuarinaceae and the Banksia– The molecular divergences among species, appropriately cor- rected for variations in the molecular evolutionary rate [63–65], Dryandra group, whereas the radiation of Mirbelieae and provide estimates of the relative time since the speciation events Bossiaeeae is somewhat later, in the early Miocene [24]. separating the species. If the rate of molecular divergence can be This suggests that the Australian radiations are even calibrated (against fossils or the occurrence of the species on islands older than the Cape radiations. – still a problematic procedure [66,67]), then the absolute divergence Although the radiations in southern Africa might have dates of the clades can be calculated. However, errors in age estimations can be large [68], so the results of dating analyses started in the Early Miocene or earlier, most modern should be treated with care. species are much more recent and have evolved during the By mapping distribution and ecological and morphological Pliocene–Pleistocene [34,35], most likely in modern attributes over phylogenetic hypotheses, the evolution of these climatic conditions (Figure 2). However, few clades have attributes can be inferred. Based on this, we can reconstruct the ecologies and distributions of the ancestors of the modern species. been sampled adequately to give an insight into the age of However, these optimizations are still problematic and we have the species. little way of knowing how accurate our reconstructions are [69]. Figure I illustrates the analytical progression from cladogram (a) What triggered the radiations? via chronogram (b) to evolutionary interpretation (c). In the Cape flora, taxonomically isolated species, or species-poor clades, are mostly restricted to forests and along permanent streams (e.g. Platylophus, Prionium and Brabejum) and to fire-protected habitats (e.g. Heeria and (a) Cladogram showing the genealogical relationships Hyenanche) [9]. These clades are largely absent from open, among the species regularly burnt, heathy vegetation or open succulent ABCDEF semi-desert habitats dominated by species belonging to the large species-radiations. This indicates that the modern species-rich clades had radiated into these summer-arid habitats when they became established in the Late Miocene, whereas the other groups represent relicts of an earlier flora [12]. In southern Africa, there is Add estimates of branch lengths, an east–west gradient in the severity of the summer corrected for variations in evolutionary rates, drought; for several groups, the greatest diversity, and and calibrate against fossils to obtain age estimates for each node. apparently the most recent radiations, are situated in the more arid west [32,34]. There are no large recent (b) Chronogram radiations in the more mesic east. This led to the ABCDEF 0 prediction that the earliest species to diverge in each 1 radiation should be found in mesic habitats, which has 2 been confirmed for Ehrharta [38] and Thamnochortus 3 [39]. For Restionaceae, the combination of anatomical and 4

Time (My) phylogenetic information also indicates that xerophytic 5 adaptations evolved relatively late in the family [40]. 6 This pattern is also seen in Australia, with the major Add ecological data for each species. Use optimization rules to infer the radiations of Acacia, Eucalyptus and Banksia into the fire- ancestral states. prone seasonally arid environments and the all-year mesic fire-protected habitats occupied by a relictual flora (c) Evolutionary interpretation: shift from wet to arid habitats, dominated by Nothofagus. The Australian fossil record is 2 My – 5 My ago, correlated with an increase in speciation rate extensive: the transformation of a mesic Nothofagus forest Arid AridArid Arid Wet Wet into a fire-prone Eucalyptus woodland during the Miocene A B CDE F 0 is well documented [41,42]. 1 Cenozoic climatic changes in the Southern Hemisphere 2 3 Change from were complex and were linked to the glaciation history of wet to arid 4 Antarctica [43], sea surface temperatures [44] and Time (My) 5 variations in the strengths of the high-pressure cells 6 over the southern oceans [45]. These factors influence both TRENDS in Plant Science the volume and seasonal distribution of rain in the Cape Floristic Region. Early Oligocene climates might have Figure I. From cladogram (a) via chronogram (b) to evolutionary interpretation (c). been similar to modern climates, with low sea-surface www.sciencedirect.com Review TRENDS in Plant Science Vol.10 No.11 November 2005 539

3200

2400 Restionaceae Pelargonium Irideae Ixioideae Indigofera Phylica Heliophila Ruschioideae

1600

Number of lineages 800

0 ) ‰ O ( 18 δ

54321 Plio.Miocene Oligocene Eocene 0 10 20 30 40 50 Mya

TRENDS in Plant Science

Figure 2. Accumulation of species diversity. The upper half of the diagram shows the accumulated retrojected estimates of diversity for those lineages for which the radiations have been dated, using the rate of lineage increase calculated for Restionaceae. Each lineage is colour-coded, and the total estimated diversity indicated by the upper line. The lower half of the graph shows the d18 O (‰) curve from marine sediments for the Eocene to present, modified from [44]. The degree of 18O/16O fractionation between seawater and carbonate ions (incorporated into the shells of marine organisms and so preserved) is influenced by the temperature. In addition,the18O/16O ratio in seawater is affected by the extent of sea-ice. The d18 O (‰) curve can be interpreted as a proxy for oceanic temperatures, and reflects the Oligocene glaciation of Antarctica, the mesic Early and Middle Miocene, and the decreasing temperatures from 14 Mya. The calculated starting dates for each radiation are indicated with colour-coded arrows, the points of which indicate the mean and the width indicates the error of the estimates. temperatures and extensive Antarctic glaciation. During innovations has been suggested: fruit and seedling the early and middle Miocene, climates were more mesic morphology in Restionaceae [50],stemanatomyin [46]. The increase in Antarctic glaciation from 14 million Aizoaceae [49] and Restionaceae [40], the annual habit years ago (Mya) [44], associated with an increase in the in the grass genus Ehrharta [38], and corms in Iridaceae strength of the South Atlantic high-pressure cell, initiated [34] and Pelargonium [32]. However, there seems to be no a trend towards the modern seasonally arid conditions. adequate test to determine whether these attributes were This was dramatically accelerated 3 Mya with the closure indeed the factors that allowed the survival and radiation of the Panama seaway [47,48]. of these lineages [51]. This rather gradual transformation in the climate means that there was no single, obvious trigger for the What drives speciation? radiation of the Cape flora. This corresponds with the Diversification in the Cape seems to be more rapid than in great spread in the dates of initiation of the radiation of other areas, at least for Restionaceae [37] and Iridaceae various lineages. Some date back to the putative arid [52], the only two clades that have been critically phase of the Early Oligocene and slowly accumulated investigated. In the past, the drivers of this rapid diversity, others started their radiation after the aridifica- diversification in the Cape flora were mainly sought in tion started 14 Mya, and some generated a large diversity mechanisms that could fragment the distribution ranges in a short time: typical recent and rapid radiations [49].In of the species and lead to geographically isolated some cases, the presence of a long basal branch in the populations, and so to allopatric speciation. Among the phylogeny (as in the Australian Daviesia [24] or the Africa postulated mechanisms were sea-level changes and Restionaceae [37]) indicates that the group was present in climatic fluctuations, as well as local extinction as a result the region for a longer period but that only one lineage of inappropriate fires [12]. from this period survives. Recently, the search for the drivers of speciation has The evolution of key innovations allowing the exploita- shifted from factors promoting allopatry to a search for tion of new seasonally arid, fire-prone habitats might have ecological parameters that could drive diversification been the crucial factor that allowed some groups to radiate between populations as well as select for reproductive while others remained species-poor and restricted to isolation [53]. Several selective forces have been proposed mesic, fire-protected habitats. A rich diversity of such for the Cape flora. Pollinator specialization, particularly www.sciencedirect.com 540 Review TRENDS in Plant Science Vol.10 No.11 November 2005 intheCapeflorawheremanyspecieshavesingle the available genetic diversity can be high. Allowing for pollinators [54–57], could lead to ethological isolation as long-distance dispersal means that with climate change well as result in strong selection for different floral additional lineages can become established in the area. morphologies. This results in reproductively isolated Species richness within this flora is the result of several species that can be recognized by their different floral radiations, some of which started in the middle Tertiary morphologies [58]. Edaphic specialization, with parapatric (Eocene–Oligocene), whereas others started only recently. sister-species restricted to the different soil types has been There has been no major extinction similar to the documented for several groups [56,59,60]. Presumably, Quaternary glaciations in Europe that massively reduced specialization to different soils results in hybrids lacking the diversity in all lineages, at least not since the suitable habitats but this has not been tested experimen- Oligocene. The final diversity of the flora is the cumulative tally. The steep climatic gradients in Southern Africa, with effect of many lineages radiating, maybe more rapidly an east–west gradient in the annual distribution of the than in other areas, over a long period of time, and rain and large orographic effects resulting in wet lee-sides possibly in response to somewhat different parameters. A and arid rain shadow interiors, might provide strong flora cannot evolve, it can only reflect the sum of disruptive selection [9,12], and numerous sister-species diversities accumulated by its evolving lineages. Environ- can be separated by their climatic preferences. In most mental heterogeneity appears to be an important factor in instances, climatic specialization also results in para- or promoting speciation and species-persistence, both allopatric distribution ranges because climatic variation important contributors to eventual high levels of species shows a geographical pattern. diversity. Final species richness might be understood as We do not know which ecological parameters are most the product of environmental diversity and the time since frequently associated with speciation in the Cape flora. the last dramatic, extinction-causing environmental The issue is complicated because we can only determine changes. which parameters are associated with species persistence: the inference that the associations we see in the present Acknowledgements day are also associated with speciation might not be This research was supported by grants from the Swiss Science justified because both distribution ranges and ecological Foundation, the Claraz Stiftung, the National Geographical Society and preferences might have changed since speciation [61]. the SANW travel grants. I thank Richard Cowling, Chloe´ Galley, Chris Hardy, Timo van der Niet and three anonymous reviewers for comments Furthermore, it seems likely that the factors promoting on the manuscript, and Alex Bernhard for preparing the figures. speciation in biotically pollinated herbs such as orchids would be different from those in wind-pollinated shrubs References such as Cliffortia. 1 Mutke, J. and Barthlott, W. (2005) Patterns of vascular plant diversity We are still far removed from a general speciation at continental to global scales. Biologiske Skrifter 55, 521–531 model for the Cape flora in which the influences of the 2 Jansson, R. (2003) Global patterns in endemism explained by past various factors are quantified. However, it seems clear climatic change. Proc. R. Soc. Lond. B. Biol. Sci. 270, 583–590 that heterogeneous environments can play an important 3 Hawkins, B.A. et al. (2003) Energy, water, and broad-scale geographic patterns of species richness. Ecology 84, 3105–3117 role. The steep selective gradients can drive rapid 4 Rosenzweig, M.L. (1995) Species Diversity in Space and Time, differentiation between populations and the diversity of Cambridge University Press habitats can provide niches for the numerous species. A 5 Davies, T.J. et al. (2004) Darwin’s abominable mystery: insights from a comparison of the otherwise similar Mediterranean floras supertree of the angiosperms. Proc. Natl. Acad. Sci. U. S. A. 101, 1904–1909 of southern Africa and western Australia suggests that 6 Cowling, R.M. and Pressey, R.L. (2001) Rapid plant diversification: the only possible explanation of the greater species planning for an evolutionary future. Proc. Natl. Acad. Sci. U. S. A. 98, richness of the African flora is the greater environmental 5452–5457 heterogeneity [11]. 7 Pressey, R.L. et al. (2003) Formulating conservation targets for biodiversity pattern and process in the Cape Floristic Region, South Africa. Biol. Conserv. 112, 99–127 Conclusion 8 Fjeldsa˚, J. and Lovett, J.C. (1997) Geographical patterns of old and Even though it has become possible to investigate the young species in African forest biota: the significance of specific evolution of plant species diversity, the generality of the montane areas as evolutionary centres. Biodivers. Conserv. 6, 325–346 interpretations are still limited by four problems. The first 9 Goldblatt, P. and Manning, J.C. (2002) Plant diversity of the Cape is that our methods of molecular dating are still flawed Region of southern Africa. Ann. Mo. Bot. Gard. 89, 281–302 10 Goldblatt, P. and Manning, J. (2000) Cape Plants. A Conspectus of the and we have little idea of how large the errors in our Cape flora of South Africa, National Botanical Institute estimations are. Second, the poor fossil record of the Cape 11 Cowling, R.M. et al. (1996) Plant diversity in Mediterranean-climate flora means that we have no independent corroboration of regions. Trends Ecol. Evol. 11, 362–366 the suggested palaeohistory. Third, the number of clades 12 Linder, H.P. (2003) The radiation of the Cape flora, southern Africa. investigated is still rather small and it is possible that the Biol. Rev. Camb. Philos. Soc. 78, 597–638 13 Levyns, M.R. (1962) Possible Antarctic elements in the South African sample was highly skewed. Finally, our methods of Flora. S. Afr. J. Sci. 58, 237–241 inferring the ecologies and distributions of ancestral 14 Adamson, R.S. (1958) The Cape as an ancient African flora. Adv. Sci. taxa are still crude. 58, 1–10 The great plant diversity of the Cape flora might be the 15 Linder, H.P. et al. (1992) History of the Cape flora. In The Ecology of Fynbos: Nutrients, Fire and Diversity (Cowling, R.M., ed.), pp. result of several processes. The incorporation of suitably 113–134, Oxford University Press adapted lineages (to the temperate climates, low-nutrient 16 Humphries, C.J. and Parenti, L.R. (1999) Cladistic Biogeography, soils and seasonal drought) from all continents means that Oxford University Press www.sciencedirect.com Review TRENDS in Plant Science Vol.10 No.11 November 2005 541

17 Givnish, T.J. et al. (1999) Polyphyly and convergent morphological 42 Kershaw, A.P. et al. (2002) A history of fire in Australia. In Flammable evolution in Commelinales and Commelinidae: evidence from rbcL Australia. The Fire Regimes and Biodiversity of a Continent (Brad- sequence data. Mol. Phylogenet. Evol. 12, 360–385 stock, R.A. et al., eds), pp. 3–25, Cambridge University Press 18 Bakker, F.T. et al. (1998) Phylogenetic relationships within Pelargo- 43 DeConto, R.M. and Pollard, D. (2003) Rapid Cenozoic glaciation of

nium sect. Peristera (Geraniaceae), inferred from nrDNA and cpDNA Antarctica induced by declining atmospheric CO2. Nature 421, 245– sequence comparisons. Plant Syst. Evol. 211, 273–287 249 19 Galley, C. and Linder, H.P. Geographical affinities of the Cape flora, 44 Zachos, J. et al. (2001) Trends, rhythms, and aberrations in global South Africa. J. Biogeogr. (in press) climate 65 Ma to present. Science 292, 686–693 20 Barlow, B. (1981) The Australian flora: its origin and evolution. In 45 Shi, N. et al. (2001) Southeast trade wind variations during the last Flora of Australia (Vol. 1) (George, A.S., ed.), pp. 25–75, Bureau of 135 kyr: evidence from pollen spectra in eastern South Atlantic Flora and Fauna sediments. Earth Planet. Sci. Lett. 187, 311–321 21 Crisci, J.V. et al. (1991) Historical biogeography of southern South 46 Udeze, C.U. and Oboh-Ikuenobe, F.E. (2005) Neogene palaeoceano- America. Syst. Zool. 40, 152–171 graphic and palaeoclimatic events inferred from palynological data: 22 Sanmartin, I. and Ronquist, F. (2004) Southern Hemisphere biogeo- Cape Basin off South Africa, ODP Leg 175. Palaeogeography graphy inferred by event-based models: plant versus animal patterns. Palaeoclimatology Palaeoecology 219, 199–223 Syst. Biol. 53, 216–243 47 Marlow, J.R. et al. (2000) Upwelling intensification as part of the 23 McGlone, M.S. et al. (2001) Endemism, species selection and the origin Pliocene–Pleistocene climate transition. Science 290, 2288–2291 and distribution of the vascular plant flora of New Zealand. 48 deMenocal, P.B. (2004) African climate change and faunal evolution J. Biogeogr. 28, 199–216 during the Pliocene–Pleistocene. Earth Planet. Sci. Lett. 220, 3–24 24 Crisp, M. et al. (2004) Radiation of the Australian flora: what can 49 Klak, C. et al. (2004) Unmatched tempo of evolution in Southern comparisons of molecular phylogenies across multiple taxa tell us African semi-desert ice plants. Nature 427, 63–65 about the evolution of diversity in present-day communities? Philos. 50 Caddick, L.R. and Linder, H.P. (2002) Evolutionary strategies for Trans. R. Soc. Lond. B Biol. Sci. 359, 1551–1571 reproduction and dispersal in African Restionaceae. Aust. J. Bot. 50, 25 Pennington, R.T. and Dick, C.W. (2004) The role of immigrants in the 339–355 assembly of the South American rainforest tree flora. Philos. Trans. R. 51 de Queiroz, A. (1998) Interpreting sister-group tests of key innovation Soc. Lond. B Biol. Sci. 359, 1611–1622 hypotheses. Syst. Biol. 47, 710–718 26 Weimarck, H. (1936) Die Verbreitung einiger Afrikanisch-montanen 52 Davies, T.J. et al. (2005) Environment, area, and diversification in the Pflanzengruppen, III-IV. Svensk Botanisk Tidskrift 30, 36–56 species-rich flowering plant family Iridaceae. Am. Nat. 166, 418–425 27 Weimarck, H. (1933) Die Verbreitung einiger Afrikanisch-montanen 53 Schluter, D. (2001) Ecology and the origin of species. Trends Ecol. Pflanzengruppen, I-II. Svensk Botanisk Tidskrift 27, 400–419 Evol. 16, 372–380 28 Carbutt, C. and Edwards, T. (2002) Cape elements on high-altitude 54 Johnson, S.D. and Steiner, K.E. (2000) Generalization versus corridors and edaphic islands: historical aspects and preliminary specialization in plant pollination systems. Trends Ecol. Evol. 15, phytogeography. Syst. Geogr. Plants 71, 1033–1061 140–143 29 Wagner, W.L. (1991) Evolution of waif floras: a comparison of the 55 Goldblatt, P. and Manning, J.C. (2000) The long-proboscid fly Hawaiian and Marquesan archipelagos. In The Unity of Evolutionary pollination system in southern Africa. Ann. Mo. Bot. Gard. 87, 146– Biology, the Proceedings of the Fourth International Congress of 170 Systematics and Evolutionary Biology (Vol. 1) (Dudley, E.C., ed.), pp. 56 Goldblatt, P. et al. (2001) Radiation of pollination systems in Gladiolus 267–284, Dioscorides Press (Iridaceae: Crocoideae) in southern Africa. Ann. Mo. Bot. Gard. 88, 30 Orchard, A.E. (1999) Introduction. In Flora of Australia. Volume 1, 713–734 Introduction (2nd edn) (Orchard, A.E., ed.), pp. 1–9, ABRS/SIRO, 57 Johnson, S.D. and Steiner, K.E. (2003) Specialized pollination systems Australia in southern Africa. S. Afr. J. Sci. 99, 345–348 31 Richardson, J.E. et al. (2001) Rapid and recent origin of species 58 Johnson, S.D. (1996) Pollination, adaptation and speciation models in richness in the Cape flora of South Africa. Nature 412, 181–183 the Cape flora of South Africa. Taxon 45, 59–66 32 Bakker, F.T. et al. (2005) Nested radiation in Cape Pelargonium. In 59 Goldblatt, P. (1982) Systematics of Freesia Klatt (Iridaceae). J. South Plant Species-level Systematics: New Perspectives on Pattern & African Bot. 48, 39–91 Process (Bakker, F.T. et al., eds), pp. 75–100, Koeltz 60 Kurzweil, H. et al. (1991) The phylogeny and evolution of the 33 Schrire, B.D. et al. (2003) Towards a phylogeny of Indigofera Pterygodium – Corycium complex (Coryciinae, Orchidaceae). Plant (Leguminosae–Papilionoideae): indentification of major clades and Syst. Evol. 175, 161–223 relative ages. In Advances in Legume Systematics, Part 10, Higher 61 Losos, J.B. and Glor, R.E. (2003) Phylogenetic comparative methods Level Systematics (Klitgaard, B.B. and Bruneau, A., eds), pp. 269–302, and the geography of speciation. Trends Ecol. Evol. 18, 220–227 Royal Botanic Gardens 62 Page, R.D.M. and Holmes, E.C. (1998) Molecular Evolution. A 34 Goldblatt, P. et al. (2002) Radiation in the Cape flora and the Phylogenetic Approach, Blackwell Science phylogeny of peacock irises Moraea (Iridaceae) based on four plastid 63 Sanderson, M.J. (1998) Estimating rate and time in molecular DNA regions. Mol. Phylogenet. Evol. 25, 341–360 phylogenies: beyond the molecular clock?. In Molecular Systematics 35 Linder, H.P. and Hardy, C.R. (2004) Evolution of the species-rich Cape of Plants II. DNA Sequencing (Soltis, D.E. et al., eds), pp. 242–264, flora. Philos. Trans. R. Soc. Lond. B Biol. Sci. 359, 1623–1632 Kluwer 36 Mummenhoff, K. et al. (2005) Phylogeny, morphological evolution, and 64 Magallo´n, S. (2004) Dating lineages: molecular and paleontological speciation of endemic Brassicaceae genera in the Cape flora of approaches to the temporal framework of clades. Int. J. Plant Sci. Southern Africa. Ann. Mo. Bot. Gard. 92, 400–424 165(Suppl.), S7–S21 37 Linder, H.P. et al. (2003) Contrasting patterns of radiation in African 65 Linder, H.P. et al. (2005) Taxon sampling effects in molecular clock and Australian Restionaceae. Evolution Int. J. Org. Evolution 57, dating: an example from the African Restionaceae. Mol. Phylogenet. 2688–2702 Evol. 35, 569–582 38 Verboom, G.A. et al. (2003) Phylogenetics of the grass genus Ehrharta 66 Conti, E. et al. (2004) Calibration of molecular clocks and the Thunb.: evidence for radiation in the summer-arid zone of the South biogeographic history of Crypteroniaceae: a reply to Moyle. Evolution African Cape. Evolution Int. J. Org. Evolution 57, 1008–1021 Int. J. Org. Evolution 58, 1874–1876 39 Linder, H.P. and Hardy, C.R. (2005) Species richness in the Cape flora: 67 Heads, M. (2005) Dating nodes on molecular phylogenies: a critique of a macroevolutionary and macroecological perspective. In Plant molecular biogeography. Cladistics 21, 62–78 Species-level Systematics: New Perspectives on Pattern & Process 68 Bell, C.D. and Donoghue, M.J. (2005) Dating the Dipsacales: (Vol. 142) (Bakker, F.T. et al., eds), pp. 47–73, Koeltz comparing models, genes, and evolutionary implications. Am. J. Bot. 40 Linder, H.P. (2000) Vicariance, climate change, anatomy and 92, 284–296 phylogeny of the Restionaceae. Bot. J. Linn. Soc. 134, 159–177 69 Hardy, C.R. and Linder, H.P. (2005) Reconstructing ancestral habitats 41 Hill, R.S. (2004) Origins of the southeastern Australian vegetation. and ecologies: accounting for intraspecific variability and issues of Philos. Trans. R. Soc. Lond. B Biol. Sci. 359, 1537–1549 timing in ecological diversification. Syst. Biol. 54, 299–316 www.sciencedirect.com